Home

Evidence of a significant genetic component to the age-related degenerative joint disease osteoarthritis has been established, but the nature of genetic influences on normal joint morphology in healthy individuals remains unclear. Following up on our previous findings on the influence of body habitus on phenotypic variation in knee joint space [Duren et al., Human Biology 78:353-364 (2006)], the objective of the current study was to estimate the heritability of radiographic joint space in the knees of healthy young adults from a community-based sample of families. A sample of 253 subjects (mean age = 18.02 years) from 87 randomly ascertained nuclear and extended families was examined. Joint width (JW) and minimum joint space in the medial (MJS) and lateral (LJS) knee compartments were measured. A maximum-likelihood variance components method was used to estimate the heritability of MJS, LJS, and JW. Covariate effects of age, sex, age-by-sex interactions, stature, weight, and BMI were simultaneously estimated. Genetic correlation analyses were then conducted to examine relationships between trait pairs. MJS, LJS, and JW were each significantly heritable (p < 0.001), with heritabilities of 0.52, 0.53, and 0.63, respectively. The genetic correlation between MJS and LJS was not significantly different from 1. Genetic correlations between each joint space measure and JW were not significantly different from 0. This study demonstrates a significant genetic component to radiographic knee joint space during young adulthood in healthy subjects. This suggests that there are specific but as yet unidentified genes that influence the morphology of healthy articular cartilage, the target tissue of osteoarthritis. Genetic correlation analyses indicate complete pleiotropy between MJS and LJS but genetic independence of joint space and JW.

Discussion

Based on studies of osteoarthritis, it is clear that genes play a significant role in cartilage health during older adulthood. What remains unclear is how much of this genetic effect is targeting cartilage integrity and how much is targeting cartilage destruction. A recent examination of 228 monozygotic twins revealed a significant genetic influence on progression of OA once established (Zhai et al. 2007). Most genetic studies of OA are designed to search for genes that are causally related to OA. However, in using this approach, one might overlook the role of genes acting before disease onset. A set of genes controlling cartilage building and/or maintenance is one example. Such a set of genes may be activated early in life and may control the attainment of peak cartilage thickness, which may then significantly affect the risk of OA later in life.

In the current study we have demonstrated a significant genetic effect (high heritability estimates) on knee joint space variability in a randomly ascertained community-based sample of healthy young adults. In our sample there is a significant sex effect on knee joint space that accounts for approximately 20% of the trait variance. We have previously shown that body size accounts for only a small proportion of the variation in joint space (Duren et al. 2006). This finding was further supported here by the covariate effects of BMI and JW, together accounting for only 3.1% of the phenotypic variance in MJS after controlling for sex and 0% of the phenotypic variance in LJS. The identification of a sex effect on joint space at this young age may help to explain the sex disparity of OA in later life.

Our study is the first to examine the genetic architecture of joint space phenotypes in a sample of healthy young adults. Compared with studies of older adults and subjects with OA, our residual heritability results are congruent with, and in some cases higher than, those reported previously. A sibling-pair study that examined knee cartilage using MRI showed that medial and lateral cartilage volumes are 65% and 77% heritable, respectively (Zhai et al. 2004). Similar results were found by Hunter et al. (2003), where femoral and tibial cartilage volumes were analyzed separately and found to be 61-76% heritable. Similarly, Spector et al. (1996) found that joint space narrowing was 41% heritable in a sample of female twins, age 48 to 70 years.

Heritabilities reported in these studies are based on populations that either have OA or are at risk for OA (e.g., menopausal or perimenopausal women). As posed earlier, it remains unclear whether the genetic architecture of OA-related traits represents a change in gene expression that maintains cartilage integrity or a change in genetic pathways that lead to degeneration, or both. The examination of populations with, or at risk of, OA obscures distinctions between these two potential genetically mediated pathways. In the current study we used only healthy young adults to elucidate the genetic effects on knee joint morphology. In so doing, we are examining the genetic underpinnings of variation in joint space before any pathway, genetic or otherwise, is deleterious to cartilage health.

In addition to a significant heritability of each joint space measure, MJS and LJS show complete pleiotropy. That is, the heritable component of joint space in both knee compartments is due to the same set of genes. In contrast, both measures of joint space show genetic independence from JW. Although these results are not unexpected, this is the first time these relationships have been empirically demonstrated. Genetic independence of joint space and joint width reinforces our previous assertion that variation in knee joint space is largely unrelated to body size.

A potential limitation of the current study is that the knee radiographs taken in the Pels Longitudinal Study, including the subset of 253 radiographs analyzed for the purpose of this study, were all taken in the traditional anteroposterior view. This technique is widely used in both clinical and research populations for the diagnosis of OA (Hart et al. 2002; Hunter et al. 2005; Manek et al. 2003), but it has been suggested that the traditional standing anteroposterior radiographie view of the knee increases variability in the appearance of joint space because of variation in knee positioning, specifically tibial plateau alignment (Vignon et al. 2003). Such an increase in stochastic variation would reduce the heritability estimate by adding nongenetic variation. Thus we consider our heritability estimates presented here to be conservative. In addition, our demonstrated high reliability and reproducibility (ICC ≥ 0.85) of joint measurements taken from the anteroposterior view indicates that findings of the current study represent the basic genetic architecture of normal variation in knee joint space.

Elucidation of the role of genes in normal healthy cartilage thickness (or its surrogate, radiographie joint space) is a timely subject. Genes coding for key components of cartilage matrix and proteins are among the prominent candidate genes identified by linkage studies of OA. These include genes such as collagen genes (COL2A1 and COL5A2), bone morphogenetic protein 2 (BMP2), prostaglandinendoperoxide synthase 2 (COX2), frizzled-related protein (FR2B), and the interleukin 2 gene cluster (Min et al. 2005; Uitterlinden et al. 2000; Valdes et al. 2004). Previous studies have identified these as OA-related genes, but each of these genes may play a key role in the development and maintenance of articular cartilage thickness. And, much like peak bone mass has been identified as the best predictor of later osteoporosis risk (Hui et al. 1999), so may we find that the attainment of peak cartilage thickness during early adulthood predicts later risk of OA. Our current findings support the role of genetics in healthy cartilage integrity and represent a step toward identifying genes involved in the maintenance and integrity of healthy cartilage tissue and risk of later OA.